The plateau pika (Ochotona curzoniae) is a keystone species in alpine meadow ecosystems on the Qinghai-Tibet Plateau. Appropriate population density is beneficial for maintaining biodiversity, whereas high density can intensify grassland degradation. To accurately understand changes in plateau pika population density and the potential risk of density-related hazard, it is necessary to develop a reliable prediction model for population density. Existing prediction models are mostly based on time-series or spatial expansion approaches, and they pay insufficient attention to the combined effects of multiple ecological factors and to their applicability across regions. Therefore, we carried out plot-based field surveys along the eastern Qinghai-Tibet Plateau, obtained active burrow density (ABD) of plateau pikas to represent population density, and simultaneously collected ecological factor data, including plant, soil, meteorological, and topographic variables. Pearson correlation analysis, principal component analysis, and the minimum data set approach were used to select explanatory variables, and an ecological factor-driven prediction model for population density was then constructed using stepwise multiple linear regression. The model was externally validated by independent data from different years and regions, corresponding to temporal and spatial scales. The results showed that plateau pika population density was significantly correlated with most ecological factors. Among the ecological factors examined, plant community coverage (Coverage), the Shannon-Wiener diversity index (H'), soil organic carbon (SOC), and mean annual temperature (MAT) were identified as key predictors. The optimal prediction model was ABD=10683.46-96.58 Coverage-324.97 H'-5.97 SOC-41.30 MAT, with a coefficient of determination (R²) of 0.88, explaining approximately 88% of the variation in population density. In addition, plant community coverage had the largest absolute standardized regression coefficient, indicating that it was the dominant predictor. In both temporal and spatial validations, relative errors between predicted and observed values at the plot scale were mostly within ±10%, demonstrating high predictive accuracy and good extrapolation ability, and supporting the use of this equation for prediction across different years and different regions within the study context. Overall, our results show that plateau pika population density is jointly regulated by vegetation structure, soil carbon pools, and regional climatic conditions, reflecting a multi-factor ecological control that integrates biotic and abiotic components. The ecological factor-driven prediction model developed here provides a quantitative basis for population monitoring and early warning, as well as for zoning management of alpine meadows, by offering an interpretable, ecology-based tool to support risk assessment and differentiated management decisions.